训练时，报错，似乎是程序没有考虑容错

[houlimin]

Traceback (most recent call last):
File "E:/TensorFlow2.0_SSD-master/train.py", line 70, in
train_step(batch_images=images, batch_labels=labels)
File "E:/TensorFlow2.0_SSD-master/train.py", line 57, in train_step
gt_boxes = gt.generate_gt_boxes()
File "E:\TensorFlow2.0_SSD-master\core\ground_truth.py", line 136, in generate_gt_boxes
gt_boxes = self.__label_positive_and_negative_predicted_boxes(valid_true_boxes, self.predict_boxes)
File "E:\TensorFlow2.0_SSD-master\core\ground_truth.py", line 98, in __label_positive_and_negative_predicted_boxes
box_true_class = box_true[..., -1]
IndexError: index -1 is out of bounds for axis 0 with size 0

[calmiLovesAI]

I will rewrite the code in the last one or two weeks.

[queliyong]

Same error +1

[jiollos]

same error, I used the rewritten code, but It still have the same problem.

[calmiLovesAI]

I have completely rewritten the code, and now the program will not report an error.